We describe some of the achievements and problems associated with prox
imal probe-based approaches to high-density data storage. While STM-ba
sed methods have demonstrated spectacular areal densities dwarfing any
thing achievable with today's storage technologies, reliability and da
ta rate issues present serious obstacles. These problems have led us t
o focus on techniques based on AFM and near-field optics. First, we ha
ve developed a thermomechanical writing scheme using an AFM tip. We ha
ve addressed many of the practical issues involved, including data rat
e. With custom low-mass cantilevers, we have demonstrated readback on
real data with a data rate of 1.2 Mb/s. We have also pursued nontopogr
aphic storage techniques based on charge storage in nitride-oxide semi
conductor structures and near-field optical storage. These techniques
should be able to achieve densities comparable to those reached with t
he AFM scheme, with the added advantage that they are fast and reversi
ble. Although it is not yet clear whether any of these probe-based app
roaches can ever be made practical, they do represent potential pathwa
ys to the higher densities that will be needed in the decades ahead.